Scientific Observer Sampling Manual ... - Kimdietrich.com

3.6.4 Random Systematic Sampling
Another way to take a random sample is to set up a random systematic frame. Random
systematic sampling can only be used when you know, or have a reasonable estimate of, the
total number of sample units. Systematic sampling involves taking a sample during every
“n th ” defined sample unit. For a random systematic frame, randomize the selection of your
first sample unit and then take every “n th ” unit thereafter. The steps for taking a random
systematic sample are the same as #1-4 above. Then,
5. Determine how many of the sample units you want in your sample.
6. Divide the total number of sample units by the number of units you want in your
sample. This gives you your value for “n”.
7. Randomly select a number between 1 and n. This will be the first sample unit in your
sample.
8. Sample every n th unit thereafter.
Example: There are 25 baskets of unsorted catch.
1. Define population – 25 baskets of catch.
2. Define sampling frame – Spatial Systematic, using baskets.
3. Define sample units – Individual baskets of fish.
4. Number all sample units – Baskets numbered as 1 – 25.
5. Determine how many sample units to sample – Decide to sample 5 baskets or 20%.
6. Calculate value of “n”: 25/5 = 5.
7. Randomly select a number between 1 and “n” – For our example this = 2.
8. Sample baskets 2, 7 (2+5), 12 (7+5), 18 (12+5) and 23 (18+5).
More examples for applying random sampling protocols to the specific gear types are in the
following chapters. The same general principles apply to selecting individual fish for
biological data collection (e.g., lengths, age structures).
3.6.5 Sample Bias
Samples can be biased in a number of ways and how you sample may increase or decrease
this bias. For instance,
• Fish stratify within trawl nets while gear is fishing. Therefore, if a sample is collected
from only one section, there’s potential for bias.
• Different species may have preferred depth distributions. When sampling fixed gear
such as longline, if samples are taken from the same depth strata all the time, then the
overall sample may become biased.
• If crew is sorting retained and discard species by hand, these fish may end up being
sorted out by size as well as species. Be aware of crew sorting bias if sampling
discarded catch after they have sorted. This is not and ideal technique but sometimes
the only way you can get a sample.
• Hand selecting fish should be avoided as the natural tendency is to select bigger or
brighter specimen.
• Small sample size can influence the variation among the samples and sometimes this
can have negative impacts on statistical analyses.
3-10